The definition of backup has changed. Backed up files or databases were typically put into some type of container, such as tar. This was always the case with backup tapes, because there was simply no way to directly transfer a file to a tape. When backups started being stored on disk, the tradition of putting backed up data into a container continued – if for no other reason than inertia. Many people feel part of what constituted a backup was this format change. However, many do not agree, including Storage Switzerland and Gartner.
For example, Gartner includes in reports entitled Critical Capabilities for Data Center Backup and Recovery Solutions and Magic Quadrant for Data Center Backup and Recovery Solutions products that store copies in their native format. Storage Switzerland believes one of the reasons that people’s understanding of backup changed is that people use backups – or simply copies of their data – differently than they used to. Customers are now starting to use copies of their data for multiple purposes, including development and testing, as well as data analytics and other data monetization projects.
This unfortunately is at odds with how traditional backups work. Data analytics aren’t possible when you store your backup in a container that requires a restore to access the data. In addition, traditional backup can only be used for dev/test by restoring something to an alternate location. Essentially, traditional backup products were really only designed to restore data; doing anything else is problematic.
But having a backup system that can accomplish multiple purposes helps to justify the cost of the system, something that can only be seen as a benefit to good data protection. This is why many “backup” products are starting to see things the other way around. They store copies of data that can be used for many things, only one of which is recovery of damaged or lost data.
Use Cases of Data Center Backup and Recovery Solutions
Where backup products are used has changed. It wasn’t that long ago one could count on backup software running in a data center located in a single place on a single type of infrastructure. Specifically, everyone used physical servers and some type of disk storage system that you could see. Now almost everything is virtual. According to Gartner’s report, backups are now found in three use cases that they call balanced physical and virtual, fully virtualized, and the public cloud. It’s interesting to note Gartner does not seem to think anyone has a purely physical environment anymore.
The first use case Gartner covers is balanced physical and virtual resources inside one or more data centers. Almost every environment has experienced virtualization to some degree, and most who have experienced virtualization have decided to expand its use within the data center. However, not all applications are easy to justify a VMware license for; therefore, there are plenty of applications that still run on physical servers. Companies with mixed physical and virtual servers need a backup product that can handle both types of environments. Purchasing, installing, and learning two different backup products increases the cost and complexity of the backup infrastructure and also increases the risk of data loss due to a mistake.
The second use case, fully virtualized, is companies that have grown so fond of the capabilities of virtualization that they decided to virtualize all applications. This may cost them more in terms of virtualization licenses, but they believe that being able to manage all servers in the same way – and back them up the same way – outweighs the associated costs. Companies that are fully virtualized should definitely select a backup product that is strong in this area, and therefore have some products available to them that those with mixed environments may not be able to use.
Finally, it is no great surprise that Gartner sees the public cloud as the final use case for backup and recovery software. When companies were using the public cloud only for development and testing, backup and recovery really wasn’t an issue. The VMs running in the cloud were simply copies of VMs on-site; there was nothing to backup. Things have definitely changed and companies are running production workloads in the cloud like never before. And contrary to popular belief, applications running in the cloud need to be backed up just as much as those running in your data center. In fact, applications running in the cloud have additional risks that you must now account for. Specifically, having your application directly accessible to the Internet makes it easier for hackers to go after your company and cause all sorts of damage. Not backing up production applications in the cloud is a really bad idea.
Critical Capabilities for Data Center Backup and Recovery Solutions
The following is a list of what Gartner considers to be the critical capabilities of a backup product for today’s applications. Storage Switzerland thinks this is an excellent list of capabilities needed in a backup product. Each is arguably quite important, but some are certainly more important than others, depending on your use case.
On-Premises Virtual Machines: This capability focuses on instant backup and recovery for virtual machines and is primarily focused on VMware. Extra credit is given for additional hypervisor support and accelerated recoveries.
User Self-Service: This capability centers on the ability for application owners to manage their own backup and recovery operations both for operational and DevOps purposes.
Storage and Network Efficiency: This area focuses on the ability to minimize storage and network requirements, and technologies like deduplication and incremental forever are considered.
On-Premises Databases/Apps: This metric focuses on breadth of application support and integration with native application tools for instant backup and recovery.
On-Premises Physical Servers: This focuses on breadth of support of physical servers and the ability to deliver fast backup, fast recovery and scalability.
Public IaaS Storage Integration: This metric focuses on a solution’s ability to integrate with object storage for long-term retention and DR, and extra points are provided for data reduction and instant recovery.
On-Premises NAS: This capability assesses the ability to perform fast backup and recoveries for NAS systems especially large ones.
Cloud Instances and Applications: This area determines the ability to protect data generated in cloud native applications like Salesforce or Office365, and it also includes production data generated in public cloud IaaS environments.
Ease of Deployment and Use: This metric focuses on ease of installation, ongoing management and the ability to test recoveries.
On-Premises Storage Integration: This capability focuses on integration with third party storage devices such as deduplication appliances, tape libraries, scale-out file systems and object storage.
The ranking for each Use Case is based on a weighted average of these 10 Critical Capabilities.
It’s rather telling that there is no “fully non-virtualized” use case. Almost every data center is using virtualization at this point, whether via traditional hypervisors such as VMware or Hyper-V, or a hypervisor running in the cloud. This creates challenges during backup, but creates a plethora of opportunities on the recovery side. Data center virtualization and VMs running in the cloud offer recovery possibilities unheard of just a few years ago. Backup and recovery products that fully leverage these possibilities will shape the face of recovery in the coming years.